[mpich-discuss] problem with execution of examples/cpi

[Ariovaldo de Souza Junior]

---------- Forwarded message ----------
From: Ariovaldo de Souza Junior <ariovaldojunior at gmail.com>
Date: 2008/6/30
Subject: Re: [MPICH2 Req #4157] Getting error message running mpiexec
To: Rajeev Thakur <thakur at mcs.anl.gov>, Pavan Balaji <balaji at mcs.anl.gov>


Hello Rajeev and Pavan,

Thanks a lot for your help on building MPICH correctly. I have removed mpd
(apt-get purge mpd) and then reinstalled MPICH. I was so happy when I saw it
was working T_T. A the looping test, with 1,000,000 loops too 431 seconds to
be performed. Do you think it is a good rate?

But now what brings me to here is what comes next. I have followed all
instructions from the "quick start" until this line:

mpiexec -n 5 examples/cpi

and it didn't work, returning this error:

problem with execution of examples/cpi  on  falcon:  [Errno 2] No such file
or directory
problem with execution of examples/cpi  on  falcon:  [Errno 2] No such file
or directory

and stopped in the second line.

Then I tried to submit this command inside of the "examples" folder, and it
returned:

problem with execution of cpi  on  eagle:  [Errno 2] No such file or
directory
problem with execution of cpi  on  falcon:  [Errno 2] No such file or
directory
problem with execution of cpi  on  falcon:  [Errno 2] No such file or
directory
problem with execution of cpi  on  eagle:  [Errno 2] No such file or
directory
problem with execution of cpi  on  eagle:  [Errno 2] No such file or
directory

Now I can see that it failed for the five processes. I have set and mounted
the nfs on the node (until now I'm setting up just one computer, later I'll
close this one) and it is working ok. So my server is named as eagle and the
node, falcon. mpd is running in both computers, this I can see utilizing the
mpdtrace command. when I submit the example "mpiexec -l -n 10 hostname" it
returns me this:

root at eagle:/modeling/mpich2-1.0.7/examples# mpiexec -l -n 10 hostname
1: falcon
2: eagle
3: falcon
5: falcon
4: eagle
7: falcon
8: eagle
6: eagle
0: eagle
9: falcon

And then I think it is ok.
the paths are set and ok, so I don't understand why is that returning that
error, not found, if even if I execute it inside of "examples" folder it
doesn't work (of course I take off the examples part, or even add the whole
path).

Ad I have a doubt, a stupid one, sorry, every time I restart the computer I
have to point the path for the bin directory, and then it works. Is there
any way to set it automatically?

Sorry for trouble you two so much. And thank you two until now for your
attention!

Ari.




2008/6/27 Rajeev Thakur <thakur at mcs.anl.gov>:

 What exact commands did you use to configure, make, and install MPICH2? How
> did music and playlist get into the mpd directory? :-)
>
> Rajeev
>
>
>  ------------------------------
> *From:* Ariovaldo de Souza Junior [mailto:ariovaldojunior at gmail.com]
> *Sent:* Friday, June 27, 2008 6:31 PM
> *To:* mpich2-maint at mcs.anl.gov
> *Cc:* mpich2-maint at mcs.anl.gov
> *Subject:* [MPICH2 Req #4157] Getting error message running mpiexec
>
> Hello everybody.
>
> I'm trying to make mpich work for a cluster running NAMD and I'm frozen in
> the tests step yet. When I try to run the simple script "helloworld.out" I
> get the followying error message:
>
> root at eagle:/modeling/mpich# mpiexec ./hello.out
> mpiexec_eagle (get_parms_from_rcfile 1403): line in mpd conf is not key=val
> pair; line=:music_directory        "/var/lib/mpd/music"
> :
> mpiexec_eagle (get_parms_from_rcfile 1403): line in mpd conf is not key=val
> pair; line=:playlist_directory    "/var/lib/mpd/playlists"
> :
> mpiexec_eagle (get_parms_from_rcfile 1403): line in mpd conf is not key=val
> pair; line=:db_file            "/var/lib/mpd/tag_cache"
> :
> mpiexec_eagle (get_parms_from_rcfile 1403): line in mpd conf is not key=val
> pair; line=:log_file        "/var/log/mpd/mpd.log"
> :
> mpiexec_eagle (get_parms_from_rcfile 1403): line in mpd conf is not key=val
> pair; line=:error_file        "/var/log/mpd/errors.log"
> :
> mpiexec_eagle (get_parms_from_rcfile 1403): line in mpd conf is not key=val
> pair; line=:pid_file        "/var/run/mpd/pid"
> :
> mpiexec_eagle (get_parms_from_rcfile 1403): line in mpd conf is not key=val
> pair; line=:state_file        "/var/lib/mpd/state"
> :
> mpiexec_eagle (get_parms_from_rcfile 1403): line in mpd conf is not key=val
> pair; line=:user                            "root"
> :
> mpiexec_eagle (get_parms_from_rcfile 1403): line in mpd conf is not key=val
> pair; line=:bind_to_address                 "localhost"
> :
> mpiexec_eagle (get_parms_from_rcfile 1403): line in mpd conf is not key=val
> pair; line=:password                        "gqtcpbm at read
> ,add,control,admin"
> :
> mpiexec_eagle (get_parms_from_rcfile 1403): line in mpd conf is not key=val
> pair; line=:filesystem_charset              "UTF-8"
> :
> mpiexec_eagle (get_parms_from_rcfile 1403): line in mpd conf is not key=val
> pair; line=:id3v1_encoding                  "UTF-8"
> :
> /modeling/mpich/bin/mpdroot: did not find secretword in mpd conf
> filempiexec_eagle (__init__ 1190): forked process failed; status=255
>
> I googled around trying all I could but I didn't find anything. Do anyone
> have a clue about what is going one? Thanks a lot for any tip!
>
> Ariovaldo
> UFAM - Brazil
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/mpich-discuss/attachments/20080630/da6f554a/attachment.htm>